Federated Learning Classification on EMNIST dir(alpha=0.1) (test)
0.063Max r_k(theta)EAGLE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EAGLElambda=1.0, Backbone=CNN2026.03 | 0.063 | -0.348 | 84.4 | 0.013 | |
| EAGLElambda=0.1, Backbone=CNN2026.03 | 0.091 | -0.506 | 84.8 | 0.029 | |
| q-FFLq=1.0, Backbone=CNN2026.03 | 0.097 | -0.473 | 83.5 | 0.028 | |
| AFLBackbone=CNN2026.03 | 0.144 | -0.497 | 84.3 | 0.036 | |
| EAGLElambda=3.0, Backbone=CNN2026.03 | 0.153 | -0.294 | 83.4 | 0.014 | |
| EAGLElambda=2.0, Backbone=CNN2026.03 | 0.178 | -0.315 | 83.8 | 0.017 | |
| EAGLElambda=5.0, Backbone=CNN2026.03 | 0.192 | -0.268 | 82.8 | 0.017 | |
| q-FFLq=3.0, Backbone=CNN2026.03 | 0.244 | -0.33 | 80 | 0.03 | |
| FedAvgBackbone=CNN2026.03 | 0.294 | -0.512 | 84.4 | 0.047 | |
| q-FFLq=5.0, Backbone=CNN2026.03 | 0.386 | -0.201 | 76.5 | 0.032 |